We celebrate two recent publications of Dr Gabriel Spadon de Souza, Dr Martha Dais Ferreira and Prof Stan Matwin from the Big Data Institute at Dalhousie University in collaboration with Dr Amilcar Soares from Memorial University:
- Unfolding AIS transmission behaviour for vessel movement modeling on noisy data leveraging machine learning
We demonstrate a way to model ship transmission of automatic identification system (AIS) messages. Ships are required to submit their metadata such as location in AIS messages, which are then captured by radio or satellite. We use the information of multiple vessels navigating simultaneously across the Atlantic and then provide the model awareness of the ocean space. We apply state-of-the-art deep learning algorithms on small and large sequences of AIS messages.
- A semi-supervised methodology for fishing activity detection using the geometry behind trajectory of multiple vessels.
We detect potential fishing activity by applying deep learning algorithms on positioning and course features within AIS messages. We discuss how the fishing activity is labelled by clustering the AIS messages. Then, we show how to use such labels to produce neural networks for fishing activity detection.